A6, Lesson 6: The future of commerce work
Tags: Sourced stories / Our view / Illustrative examples. Self-paced reading and practice. No math or code. Every case or news claim carries its source and the date we read it; invented figures are labeled invented.
1. The story
In 2012, Amazon bought a small robotics company called Kiva Systems and began filling its warehouses with squat orange robots that carry shelves to human workers. By 2022, Amazon's own count was more than 520,000 robotic drive units rolling through its buildings. The surprise in the same company's retrospective: over that same decade, Amazon added more than a million jobs. The robots grew. The jobs grew. The work changed shape.
2. The one idea
Tasks change before jobs vanish. The warehouse worker who once walked miles each day now works alongside robots and handles what they cannot - the damaged box, the mislabeled shelf, the exception. One company's decade is not a law of economics, and other industries may not follow it. But the pattern to watch is tasks, not job titles.
3. The kitchen-table version
The dishwasher in a family restaurant. Nobody's job was "carry hot water." The machine took the rinsing, and the person it freed up ended up plating, expediting, tasting - the parts with judgment. The job title stayed; the afternoon changed.
4. Worked example
Our invented two-person candle shop lists its week honestly. Drafting thirty product descriptions: draftable. First replies to "where is my order": draftable. Deciding whether to drop a supplier who keeps missing shipments: judgment. Saving the wholesale account after a late delivery: judgment. Invented, script-checked figures: if the draftable list takes 14 of their 80 combined weekly hours, and AI drafts plus human checking cut that to 6, the shop frees 8 hours a week - a full working day - without a new hire. That day goes to the judgment list, where the growth actually lives. Six months later the same two people run a bigger shop - same job titles, different Tuesday afternoons.
Instructor: "Why not automate the judgment list too, eventually?" Learner: "Because the judgment list is where the shop's reputation compounds. Dropping a supplier or saving an account is exactly where being a person who knows the history is the product."
5. Why this matters to a shop
The split between "AI drafts, human checks" and "human owns" is arriving in small commerce now, in listings, support, and marketing. Owners who sort their tasks on purpose keep the savings and the soul. Owners who drift lose the savings to tool sprawl and the soul to generic copy.
6. What it costs and what it breaks
Every tool adds a subscription, a learning curve, and a new thing to check. Over-automating customer contact can cost more in loyalty than it saves in hours. And the "human checks" half is real work - skipping it is how drafts full of confident errors ship under your name. The freed day is only real if somebody's calendar actually changes.
7. Our habits, not rules
List the tasks your week actually contains, then sort them into "AI drafts, human checks" and "human owns." Hire and train for the second list. When you adopt a tool, redesign the job around it on purpose, and keep the checking step staffed.
8. Monday checklist
Write down last week's tasks, honestly, to the quarter hour. Mark each D (draftable) or J (judgment). Put a rough hour count on the D list. Choose one D task for an AI-draft-plus-human-check trial. Block the freed hours on a calendar for a named J task, or the savings will evaporate. Decide who checks the drafts, and when.
9. Try it yourself (five minutes)
Write down everything you did yesterday, honestly, to the quarter hour. Mark each task D for draftable or J for judgment. The J list is your future job description, and probably your moat.
10. Self-check
- What grew at Amazon alongside the robots, per its own retrospective? Answer: More than a million jobs over the same decade that drive units passed 520,000 - one company's count, not a neutral forecast.
- Why does that not guarantee the same outcome elsewhere? Answer: One company's decade is not a law of economics; watch tasks, not job titles.
- In the worked example, where do the freed eight hours come from? Answer: Fourteen draftable hours cut to six by AI drafts plus human checking (invented figures).
- Which task on your list would you hand to AI first, and which would you never hand over?
- Who does the checking in your shop, and is that time budgeted?
Sources
- Amazon, "10 years of Amazon robotics," 2022: Kiva acquired 2012, 520,000+ robotic drive units, more than a million jobs added over the decade - one company's retrospective, not a neutral forecast; figures are as of that 2022 post, read 3 Nov 2026.
- Wick and Ember's week, the 14-to-6-hour split and the eight freed hours are invented, illustrative and script-checked, not a reported story.
- Review flags: "tasks change before jobs vanish" is our framing of one company's story, not an economic finding.
- Go deeper, free: Harvard's session 10, The Future of Work, compares the optimistic and cautious forecasts and looks at which tasks are exposed first. Watch it here.
Review flags
- "tasks change before jobs vanish" is our framing of one company's story, not an economic finding.
Claim ledger
| Claim or example | Basis | Checked |
|---|---|---|
| Amazon says Kiva acquired in 2012; 520,000+ drive units and 1M+ jobs over the decade by 2022. | Amazon retrospective, 2022 | 3 Nov 2026 |
| One company retrospective, not an economic forecast; shop hours are invented. | Invented, illustrative; arithmetic script-checked | 3 Nov 2026 |
Independence
Independently developed by Commerce Stories, not affiliated with Harvard. We link to the archived 2024 course; we do not reproduce its videos, slides or exercises. The linked videos were not watched for this written-material review.